AI Can Generate the Slides. We Still Need to Generate the Insight

AI can create a polished PowerPoint presentation in minutes.

But can it explain the reasoning behind every recommendation on the slide?

That question changed the way I think about using Copilot.

This Sunday, I was reflecting on one of the goals in my Individual Development Plan: becoming a stronger stakeholder and contributing to better decisions, not just better outputs.

While thinking about that, I came across an article by Paul Millerd, How to use AI in knowledge work and consulting work without shipping slop.

What I expected was another article about prompting techniques and productivity.

What I got instead was a challenge to the way I use AI.

Like many of us, I use Copilot regularly, especially in PowerPoint. It helps me structure ideas, summarize information, build storylines, and create first drafts much faster than before.

And honestly, I often look at the result and think:

"That's pretty good."

But one example from the article made me pause.

Millerd asked an AI model to create a slide about U.S. real estate trends. The first result looked impressive: charts, headlines, and a convincing storyline. At first glance, it looked ready for an executive presentation.

Then he started asking simple questions:

• What exactly do we mean by "rebound"?
• Do the data actually support the headline?
• Is the title aligned with the evidence?
• What's the real takeaway?
• Why should an executive care?

With each round of questioning, the slide improved. The insight became sharper. The story became clearer.

The lesson was simple: the first draft looked finished, but the thinking wasn't.

One point from the article stayed with me:

AI lowers the cost of producing work, but increases the value of judgment, understanding, and synthesis.

That example felt very familiar.

How often do we stop at the first polished output because it looks convincing?

How often do we confuse a well-designed slide with a well-developed idea?

The more I reflected on it, the more I realized that I sometimes use Copilot as a content generator when I should be using it as a thinking partner.

Not to tell me what to think.

But to challenge my thinking.

Instead of asking:

"Create a slide for me."

Maybe the better questions are:

"What assumptions am I making?"

"What am I missing?"

"How would a skeptical stakeholder challenge this?"

"Would I feel comfortable defending this conclusion in front of my leadership team?"

That's where the real value starts.

Because as AI becomes better at producing content, our value shifts from creating slides to creating clarity.

From generating answers to asking better questions.

From presenting information to generating insight.

My biggest takeaway from this Sunday reflection:

The first AI output should be the start of the conversation, not the end of the thinking.

Key Takeaways

• AI can generate content, but humans still need to generate insight.
• A polished presentation is not the same as a well-developed recommendation.
• The value of knowledge workers is shifting from content creation to judgment and synthesis.
• Copilot works best as a thinking partner, not just a content generator.
• Better questions often create more value than faster answers.

Final Thought

The most important skill in the age of AI may not be prompting.

It may be judgment.

As tools like Microsoft Copilot become better at generating content, the competitive advantage of knowledge workers will increasingly come from framing problems, challenging assumptions, and turning information into actionable insights.

AI can create the presentation.

People still need to create the understanding.

📖 Inspired by Paul Millerd's article: How to use AI in knowledge work and consulting work without shipping slop
https://strategyu.co/using-ai-in-knowledge-work-and-consulting/

Has AI mainly changed the way you work, or has it also changed the way you think?

#AI #Copilot #KnowledgeWork #StrategicThinking #Learning #FutureOfWork